Population Ecology of the Barnacle Chtliamalus Stellatus in the Northwest Mediterranean

Home , Chthamalus stellatus, Semibalanus balanoides

MARINE ECOLOGY PROGRESS SERIES Published June 5 Mar Ecol Prog Ser I

Population ecology of the barnacle Chtliamalus stellatus in the northwest Mediterranean

L. Benedetti-Cecchi*, S. Acunto, F. Bulleri, F. Cinelli

Dipartimento di Scienze dell'uomo e delllAmbiente, via A. Volta 6,56126 Pisa, Italy

ABSTRACT: This study examined patterns in the distnbution and demography of the barnacle Chthamalus steiiatus (Poli) at different spatial scales in the northwest Mediterranem. Prelirninary data indicated that the abundance and size of barnacles decreased from high-shore to low-shore habitats. The generality of these patterns was investigated at several locations (10s to 100s of km apart), at several sites within locations (100s to 1000s of m apart) and at different times. Patterns were consistent with the prelirninary observations, despite considerable spatial and temporal variabiiity at smaii and large spatial scales. The foliowing models were proposed to explain the observed patterns: (1)recruitment was intnnsicaiiy greater high on the shore, (2) limitation of recruitment due to pre-emption of the substratum was greater low than high on the shore, (3) environmental conditions reduce growth low on the shore, and (4) mortality was greater low on the shore. The predictions of these models were tested by examining patterns of recruitment, growth and mortality of barnacles and avaiiability of free space in relation to height on the shore, at several spatial scales and through time. Successful recruitment of barnacles was observed at different heights on the shore where resident organisms were removed, despite a trend toward a larger number of recruits high on the shore at 1 location (Livorno). Availabii- ity of bare rock for recruitment was greater high on the shore, implying that pre-emption of the substratum was more intense low on the shore. There was no evidence to suggest that barnacles grew faster on the high shore than on the low shore. The opposite pattern was observed for young barnacles in several cases. Mortality rates were generally greater in low-shore than high-shore habitats for young and for adult barnacles. Patterns emerged over a background of considerable spatial and temporal vanation. These results emphasized the importance of pre-emption of space and mortality of juveniles in generating patterns in the distnbution and structure of populations of barnacles on rocky shores in the northwest Mediterranean.

KEY WORDS: Barnacles . Chthamalus steiiatus . Demography . Multiscaie analyses . Spatial vanation . Temporal vanation

INTRODUCTION fore, investigating the magnitude, scales and causes of demographic variation is central to the development of As an ultimate goal ecology should provide explana- ecological theory. The importance of scale in the tions for the variability observed in the distribution analysis of pattern has recently received considerable and abundance of natural populations. Understanding attention (Levin 1992, Schneider 1994), and scales of causal processes initially relies on appropriate de- variation in the distribution and abundance of popula- scription of pattern. Manipulative experiments are tions have been documented both in terrestrial and then necessary to identify the mechanisms underlying aquatic systems (Carlile et al. 1989, Rossi et al. 1992, the observed patterns. Patterns in assemblages are Thrush et al. 1994, Underwood & Chapman 1996). generated by physical and biological processes influ- Barnacles are convenient organisms to address these encing the recruitment, growth and mortality of organ- issues. They are abundant and common on rocky isms at a variety of spatial and temporal scales. There- shores throughout the world (Stephenson & Stephen- son 1949), their biology and life histories are well known (Southward 1987), and they have been used as test organisms to investigate ecological models related

O Inter-Research 2000 Resale of fuii article not permitted 158 Mar Ecol Prog Ser

to competition and predation (Connell 1961a,b), dis- low on the shore because environmental conditions persal and recruitment (Grosberg 1982, Caffey 1985, reduce growth in comparison to high-shore habitats Gaines & Roughgarden 1985, Roughgarden et al. 1987, (perhaps due to interactions with other organisms), Raimondi 1990), and demographic models for species and (4) barnacles are smaller low on the shore because with complex life cycles (Roughgarden et al. 1985, environmental factors increase mortality so that they 1988). More recently, studies have investigated the have a smaller probability of growing to adult Stages roles of water flow, supply of food and availability of than do those occurring higher up on the shore. The free space on the dynamics and structure of popula- first model predicts that if patches of bare rock are tions of intertidal barnacles (Gaines & Roughgarden cleared at different heights on the shore (to reduce the 1985, Roughgarden et al. 1985, Bertness et al. 1991, effects of resident organisms), then recruitment wiil be Minchinton & Scheibling 1993, Sanford et al. 1994). greater in high-shore than low-shore habitats. The sec- Although some of these studies have included analyses ond model predicts that bare rock available for recruit- of scales of variability in patterns of distribution and ment wiii be greater low on the shore rather than abundance, multiscale analyses of demographic vari- higher up. The third model predicts that growth of bar- ables are uncommon (but See Bertness et al. 1991, nacles will be faster high on the shore, while the fourth Hyder et al. 1998). model predicts faster rates of mortality low on the The barnacle Chthamalus stellatus (Poli) is the most shore. Clearly, these models are not mutualiy exciusive common sessile invertebrate in midlittoral rocky shore and the observed patterns in abundance and size of assemblages of the northwest Mediterranean. These barnacles may result from any possible combination of organisms may occur at various heights on the shore the proposed explanations. but, on average, are more abundant in high-shore To discriminate among the alternative explanations, habitats (0.4 to 0.7 m above Mean Low Water Level, we examined of recruitment of barnacles in MLWL), despite considerable variation in density of experimental plots cleared at different heights on the barnacles among sites (Benedetti-Cecchi et al. 1999, shore, in different places and at different times. In Menconi et al. 1999). Previous experimental studies addition, annual growth and mortality were exarnined have shown that the recruitment of barnacles low on for 2 consecutive years at different spatial scales, rang- the shore may be limited by the availability of free ing from patches a few metres apart to locations hun- space and by interactions with macroalgae (Benedetti- dreds of kilometres apart. Cecchi et al. 1999, Benedetti-Cecchi 2000). There is also a tendency for low shore barnacles to be smailer and have a lower reproductive output than individuals MATERIALS AND METHODS higher on the shore (Benedetti-Cecchi & Acunto unpubl. data). These patterns, however, have been docu- Study system. This study was done at 4 localities in mented at a limited number of sites and are not indica- the northwest Mediterranean, 10s to 100s of km apart tive of any general regional Pattern. (Fig. 1).These localities supported similar assemblages Here, we have investigated scales of variability in of algae and invertebrates, although the relative abun- the distribution and demography of Chthamalus steila- dance of some organisms and their specific patterns of tus at a number of locations in the northwest Mediter- distnbution differ among localities (Menconi et al. ranean. C. montagui was not considered because it 1999, Benedetti-Cecchi et al. unpubl. data). Barnacles was uncommon on the shores examined, although were the most common sessile invertebrates in mid- rocky shores in the northern Part of the Ligurian Sea and high-shore habitats. The upper limit of Chtham- and the Adriatic can support dense populations of this alus stellatus was between 30 and 40 cm above the species (Pannacciulli et al. unpubl. data). First, we upper limit of the tides at our study sites. At mid tested whether the preliminary observations that bar- heights, barnacles were rnixed with macroalgae, par- nacles higher on the shore were more abundant and ticularly the fleshy red alga Rissoeila verruculosa larger than those occurring lower down were consis- (Bertolini) J. Agardh, filamentous green and red algae, tent at different spatial scales and through time. Then the encrusting brown Ralfsia verrucosa (Areschough)

we tested the predictions of a number of models that ' J. Agardh and cyanobacteria (genera Calothxix and could explain the observed patterns: (1) barnacles are Rivularia). Low on the shore, encrusting coralline more abundant high on the shore because recruitment algae were common and barnacles more sparse. is intrinsically greater there than low on the shore (per- Limpets Patella aspera Roeding and P. rustica L. are haps due to physical transportation of larvae), (2) bar- the most important grazers on these shores while car- nacles are more abundant high on the shore because nivores, such as the whelk Thais haemastoma (L.) and recruitment is prevented by pre-emption of the sub- the sea Star Coscinasterias tenuispina (Lamark), were Stratum in low-shore habitats, (3) barnacles are smailer occasionally present at some localities. Benedetti-Cecchi et al.: Population ecology of barnacles 159

Patterns in size and abundance of barnacles and Data on barnacle size were analysed using a 5-factor availability of bare rock. To test the hypothesis that mixed model ANOVA with the foilowing factors: Loca- barnacles high on the shore were consistently larger tion (random and orthogonal), Site (random, nested than those lower down (as initially observed at Livorno; within Location), Patch (random, nested within Site), Benedetti-Cecchi & Acunto unpubl. data), we com- Height (fixed and orthogonal) and Plot (random and pared the size of barnacles at 2 heights on the shore at a nested within the interaction between Height and hierarchy of spatial scales in the northwest Mediter- Patch). A 5-factor mixed model ANOVA was also used ranean. The sampling design included 3 locations 100s to analyse patterns of abundance of barnacles and of km apart (Baratti, Capraia and Livorno, Fig. I),3 sites availability of bare rock. Factors were: Location (ran- within each location (100s to 1000s m apart) and 3 dom and orthogonal). Date (random and nested within patches within each site (3 to 10 m apart). A site was a Location), Site (random and nested within Date), Patch stretch of coastline 30 to 50 m in length, while patches (random and nested within Site),and Height (fixed and were 2 to 3 m in length. In July-August 1996,6replicate plots (7 X 5 cm) were photographed at high (40 to 50 cm above MLWL ) and low (10 to 20 cm above MLWL) heights in each patch, using a Nikonos V carnera with an electronic flash and an extension tube for macro photography (l:2 reproduction ratio with a 35 mm lens). The Same sites were used for the analysis of patterns of growth and mortality of barnacles descnbed below (see 'Growth and mortality'). Five barnacles were selected randornly in each slide and their size recorded as the opercular diameter measured under a dissecting microscope with a rnicrometric eye piece (to the near- est 0.01 mm). The distnbution of barnacles and the availability of bare rock were examined at 2 localities (Capraia and Livorno, Fig. 1). Each location was sampled twice between January and May 1998. For each date, 6 sites (stretches of shore of 10 to 15 m) were selected randomly along 4 to 5 km of rocky coast at each location. Two patches (stretches of shore of 1.5 to 2 m) were selected randomly at each site and the percentage cover of barnacles and bare rock was assessed at 4 heights in each patch, by sampling 3 replicate quadrats of 15 X 10 cm at each height. Heights were at levels on the shore 30, 4 km 20, 10 and 0 cm above the MLWL, as determined from tide-tables. These heights matched macroscopic changes in the structure of assemblages along the vertical gradient. Percentage cover values were determined using a plastic frame divided into 25 sub-quadrats of 3 X 2 cm each, and after visual inspection a Score from 0 to 4% was given to each taxon in each sub-quadrat. Final cover was obtained by surnrning over the 25 smali quadrats (Dethier et al. 1993, Benedetti- Cecchi et al. 1996). Fig. 1. Map of the study area in the northwest Mediterranean 160 ' Mar Ecol Prog Ser 198: 157-170, 2000

orthogonal). In this and the subsequent analyses, Plot (i.e.residual vanation), Patch and Location, calcu- Cochran's C-test was used to examine the assumption lated. Observed values were then equated to the ex- of homogeneity of vanances. If necessary, data. were pected mean squares to obtain estimates of spatial . log-transformed to eliminate or reduce heterogeneity vanance at the different scales (see Underwood 1997 of vanances. Student-Neuman-Keuls (SNK) tests were on how to denve expected mean squares for any given used for multiple comparisons of the means (at a = design). 0.05). Anderson's test (Winer et al. 1991) was used to Growth and mortality. Annual growth and mortality test for the significance of differences in the frequency of barnacles were examined at the Same hierarchy of at which treatments were given a particular rank by spatial scales at which the initial size of barnacles was SNK tests in a senes of niultiple compansons. measured, but in 2 consecutive years (1996-1997 and Patterns of recruitment. Patterns of recruitment 1997-1998). Permanent quadrats were established at were examined in plots of 10 X 15 cm scraped clean of each of 3 sites at each location (Baratti, Capraia and resident organisms at Livorno and Punta Bianca Livorno). Six patches of substratum were randomly (Fig. 1). The experiment was initiated before the main selected and marked with epoxy putty at each site in settlement penod of barnacles, which occurs in July- July-August 1996. To obtain temporally independent August at these locations (Benedetti-Cecchi & Acunto estimates of growth and mortality, 3 patches (selected unpubl. data). A first Set of plots was cleared in Janu- randomly among the 6 avaiiable) were used in the iirst ary 1998 and a second set was cleared in June 1998 at year of the study. The other 3 patches were used in the each location. In this way, plots of different age with a second year. At the beginning of each penod, 5 repli- different degree of free space (Benedetti-Cecchi & cate plots were marked at each of 2 heights on the Bertocci unpubl. data) were available for barnacles to shore in the appropnate patches. Plots were 7 X 5 cm in settle, allowing us to test whether recruitment is af- size and were delimited by 2 holes drilled into the rock fected by the history of the plot (i.e. the total effect of that corresponded to the upper corners of the quadrats. availability of free space and composition of resident Smaii pieces of epoxy putty were also used to facilitate assemblages, both in and surrounding the cleared the identification of quadrats. The 2 heights corre- plots). For each date at each location, 2 replicate sponded to high-shore barnacles (40 to 50 cm above patches (stretches of shore 3 to 4 m in length) were MLWL) and to low-shore barnacles (10 to 20 cm above selected randornly at each of 4 heights on the shore. MLWL). Three replicate quadrats in each patch were scraped Plots were photographed as descnbed above (see clean using a hammer and chisel, and marked at their 'Patterns in size and abundance of barnacles and avail- corners with epoxy putty (Subcoat S, Veneziani) for ability of bare rock'). The holes dnlled into the rock relocation. The 4 heights chosen were: (1) above the facilitated the positioning of the frame so that the Same upper limit of distribution of Rissoelia verruculosa (35 area was sampled at the beginning of the study and to 40 cm above MLWL); (2) at the upper margin of dis- after 1 yr. Growth was measured as the annual in- tribution of R. verruculosa (25 to 30 cm above MLWL); crease in opercular diameter for juvenile (<1.5 rnrn (3) at the lower margin of distnbution of R. verruculosa opercular diameter) and adult (>2.5 rnrn opercular (15 to 20 cm above MLWL), and (4) below the lower diameter) barnacles. Five barnacles of each size class limit of distnbution of R. verruculosa (5 to 10 cm above were selected from each slide from the first time of MLWL). sampling, and their opercular diameter measured. Recruitment was monitored in August 1998 at both These barnacles were selected randomly with the pro- locations. Plots were photographed using a Nikonos V viso that they survived until the next year, when they camera equipped with electronic flach and an exten- were measured again. Annual growth was calculated sion tube for macro photography (1:3 reproduction ratio as the differente between final and initial measure- with a 35 mm lens). Slides were examined under a dis- ments and averaged across the 5 barnacles in each secting microscope and the number of Spat counted. plot. Similarly, 5 barnacles of each size class were Data were analysed using a 4-factor ANOVA with the tracked over the 1 yr penods to calculate the percent- following treatments: Location (random), Age (fixed age mortality for each plot. These data were analysed and orthogonal to location), Height (fixed and orthogo- using a 5-factor ANOVA, with the following factors: nal to Location and Age) and Patch (random and Time (random), Location (random and orthogonal to nested within the Height X Location X Age interaction). Time), Site (random, nested within location but ortho- To further examine vanation in recruitment, esti- gonal to Time), Patch (random and nested within the mates of spatial variance at different scales were cal- Time X Site interaction) and Height (fixed and ortho- culated from the data. In this case, ANOVAs were gonal to Time, Location, Site and Patch). Some plots repeated independently for each age and height on the were missing by the end of the 1 yr period. This shore and the mean squares, associated to the effect.of occurred either because plots could not be relocated, Benedetti-Cecchi et al.: Population ecology of barnacles 161

or because of proliferation of algae that masked the as indicated by the significant Height X Patch and barnacles present in the plots. Missing values were Height X Location interactions (Table I), SNK tests replaced with mean values calculated from the re- within these interactions indicated that barnacles high maining plots in each particular patch, and the degrees on the shore were always significantly larger than of freedom of the residuals adjusted accordingly those occurring lower down (27 comparisons within (Underwood 1997). Estimates of spatial and temporal the Height X Patch interaction and 3 comparisons vanance in growth and mortality were obtained as de- within the Height X Location interaction). The average scnbed above (see 'Patterns of recruitment') for juve- magnitude of these differences was 34.5 % at Baratti, nile and adult barnacles. 21.7% at Capraia and 48% at Livorno. There was, however, considerable variation in the mean size of barnacles at the scale of 10s to 100s of cm, irrespective RESULTS of height on the shore, as indicated by the significant effect of Plot in the analysis (Table 1). An example of Patterns in size and abundante of barnacles and these patterns is iüustrated for barnacles at Capraia availability of bare rock (Fig. 2). There were inconsistencies in the vertical distribu- Barnacles were significantly larger high than low on tion of barnacles from patch to patch, site to site and the shore, consistently over the spatial scales exam- between locations. These patterns resulted in signifi- ined. Although the magnitude of the differences be- cant interactions between Height and each of the 3 tween heights changed across patches and localities, spatial scales examined (Table 2). This analysis, how-

Table 1. Chthlamalus steilatus. ANOVA of mean sizes. Significant effects relevant to the interpretation of the results are indicated in bold. L: Location, H: Height

1 Source of variation df MS F P Denominator for F I L 2 5.958 Site (L) Site (L) 6 2.906 Patch (Site (L)) Patch (Site (L)) 18 2.247 Plot (Patch (Site (L))) H 1 236.262 HxL HxL 2 10.217 H X Site (L) H X Site (L) 6 0.857 H X Patch (Site (L)) H X Patch (Site (L)) 18 0.999 Plot (H X Patch (Site (L))) Plot (H X Patch (Site (L))) 270 0.420 Residual Residual 1296 0.616 Cochran's C-test C = 0.0165,p > 0.05 Transformation None

Table 2. Chthamalus steilatus. ANOVA of mean percentage Covers. Significant effects relevant to the interpretation of the results are indicated in bold

Source of variation df Cover of C. steilatus Bare rock Denominator for F MS F P MS F P

L 1 52.04 5.42 >0.1 7831.6 2.2 >0.25 Date (L) Date (L) 2 9.60 2.41 >0.1 3557.1 4.1 <0.05 Site (Date (L)) Site (Date (L)) 20 3.99 4.98 <0.0002 867.9 4.4 <0.0005 Patch (Site (Date (L))) Patch (Site (Date (L))) 24 0.80 2.88 <0.0001 197.4 1.7 <0.05 Residual H 3 198.60 7.52 >0.05 11791.7 50.4 <0.005 H X L HxL 3 26.41 13.45 <0.005 234.1 0.3 >0.8 H X Date (L) H X Date (L) 6 1.96 1.88 >0.09 820.7 1.3 >0.25 H X Site (Date (L)) H X Site (Date (L)) 60 1.04 2.06 <0.002 629.8 2.8 ~0.0001 H X Patch (Site (Date (L))) H X Patch (Site (Date (L))) 72 0.51 1.82 <0.0003 222.9 1.96 ~0.0001 Residual Residual 384 0.28 Cochran's C-test C = 0.1155,p < 0.01 C = 0.0562,p > 0.05 Transformation Ln(x + 1) None 162 Mar Ecol Prog Ser 198: 157-170, 2000

High shore 0Iw shore Patch I Site 1 Site 2 Site 3 41 41 41

8 C) o> Patch 2

Patch 3 41

Plot number

Fig. 2. Chthamalus stellatus. Illustrative example of mean opercular diameter (+SE, n = 5) at 3 sites, 3 patches within each site and 2 heights on the shore at Capraia. On average, barnacles high on the shore were 21.7 % larger than those occurring lower down. The magnitude of these differences was 34.5 % at Baratti and 48 % at Livorno (data not shown; See text for details) .

ever, must be interpreted cautiously because of hetero- SNK and Anderson's tests indicated that 30 cm above geneity of variances. SNK tests within the Height X. the MLWL ranked first in the availability of free space Patch interaction indicated that barnacles were signif- significantly more often than any other height (Height icantly more abundant 30 cm above the MLWL than at' X Patch interaction: Q2 = 16.4, p < 0.005; Height X Site any other height, consistently across patches (Ander: hteraction: Q2 = 13.6, p < 0.005). son's test: Q2 = 36.8, df = 3, p < 0.001). Thus, the interaction resulted from spatial variation in the magnitude of these differences rather than from changes in the Patterns of recruitment rank order of the heights. An example of these Patterns is iiiustrated for 6 sites sampled on Date 1 at Livorno Recruitment was massive at Livorno, while few spat (Fig. 3).A similar Pattern of differences arnong heights were present in quadrats at Punta Bianca (Fig. 5). The was obseived at Livorno on Date 2 and at Capraia. analysis detected a significant Height X Location inter- Bare rock avaiiable for recruitment was greater high aition (Table 3). SNK tests within this interaction indi- on the shore rather than lower down both at Livorno cated that there was no significant difference among and Punta Bianca (Fig. 4). Patterns were, however, heights at Punta Bianca, and that the only difference variable from patch to patch and site to site aslindi- occurring at Livorno was that between the 2 rnost cated by the significant Height X Patch and Height X, extreme heights considered (i.e. there were more Site interactions (Table 2). Despite these interactions, recruits in plots 35 to 40 cm above the MLWL than in Benedetti-Cecchi et al.: Population ecology of barnacles 163

sistent across patches and also interacted with time and location (Table 5). Anderson's test within the Height X Patch interaction indicated that the mean 20 cm above MLWL growth of barnacles living down the shore was 10 cm abve MLWL OcmaboveMLWL ranked highest significantly more often than the 40 mean growth of those occurring high on the shore (Q2= 13, df = 1, p 0.001). SNK tests within the 20 Height X Time X Location interaction indicated that barnacles in low-shore habitats had higher growth O Patch I Patch 2 Patch 1 Patch 2 rates than those living up on the shore at Capraia in Year 1 (Fig. 6B), and at Livorno in Year 2 (Fig. 6C). In none of these cornparisons was there evidence for barnacles living high on the shore growing faster than those living lower down. Vanance in growth of barnacles was largest at the scale of the location and then at the scale of the plot, 20 while vanability among patches and sites was negligible (Table 6). Growth was more vanable in adult Patch 2 Patch 2 LO Patch I than juvenile barnacles at the scale of the location, L both in high-shore and low-shore habitats. Con- E Site 5 versely, at the scale of the plot, growth was more

'O0IRO.. '0°180 F I vanable in juvenile than adult barnacles. There was no temporal variability in Patterns of growth of barnacles (Table 6).

40 Mortaiity of juvenile barnacles differed between heights, but the differences were not consistent 20 across patches and changed interactively with time 60[L1 and location (Table 7). Nevertheless, Anderson's O Patch l Patch 2 Patch 1 Patch 2 test within the Height X Patch interaction indicated that mortality of low-shore barnacles was ranked Fig. 3. Chthamalus stellatus. Mean percentage cover (+SE, highest significantly rnore often than ranks for high- n = 3) at different heiqhts on the shore. The example iiius- trates data from 6 sitesat Livorno on Date 1. A similai pattern of differences among heights was observed at Livorno on A Livorno Date 2 and at Capraia (data not shown; See text for details)

40 date 40 date plots 5 to 10 cm above the MLWL). However, the SNK test could not identify a clear alternative to the null hypothesis of no difference among - ,o heights.Spatial variance in recruitment was very large at W$Q; l.LLLo 30 20 10 o 30 20 10 o > the smallest spatial scale (among plots) in low- Q B Punta Bianca shore habitats, and decreased with increasing date date 2 height on the shore (Table 4). The magnitude of 4 so, vanation among plots was comparable to that observed between locations in high-shore habitats (20 and 30 cm above the MLWL). Variability among patches was negligible (Table 4). ::L2010 /L 0 Growth and mortality of barnacles 30 20 10 0 30 20 10 0 Height above MLWL (cm)

There differentes in Of were Patterns growth of Fig. 4. Bare rock. Mean percentage cover (+SE,n = 36) at different Juvenile barnacles between high-shore arid low- heights on the shore. Data are from 3 repiicate plots pooled across 2 shore habitats. The effect of height was not con- replicate patches and 6 repiicate sites on each date at each location 164 Mar Ecol Prog Ser 198: 157-170, 2000

Table 3. Chthamalus steiiatus. ANOVA of mean densities of recruits. Significant effects relevant to the interpretation of the results are indicated in bold. A: Age

I Source of variation df MS F P Denominator for F I

L Patch (H X L X A) A AxL H HxL AxL Patch (H X L X A) HxL Patch (H X L X A) HxA HxLxA HxLxA Patch (H X L X A) Patch (H X L X A) Residual Residual Cochran's C-test C = 0.1709, p > 0.05 Transformation Ln(x + 1)

...... Table 4. Estirnates of spatial vanance in recruitment of barna- A Livomo cles at different scales

Age Height on oZplot 02patch 02LocaDon (yr) the shore (cm)

1 0 2319.5 107.0 648.7 1 10 2376.0 50.3 359.0 1 20 180.1 0.0 144.0 1 30 301.1 0.0 36.3 2 0 1555.0 0.0 36.3 0 2 10 74.6 0.0 51.5 .M 80- 8 B Punta Bianca 2 20 54.3 0.0 46.7 U 2 30 38.5 0.0 13.1 n 60- tion indicated faster rates of mortality of low-shore barnacles in 4 out of 6 compansons. In none of these compansons were rates of mortality of high-shore barnacles significantly larger than those occurnng lower ''35-40 25-30 15-10 5-10 down. Height above MLWL (cm) Mortality of barnacles was very variable at the scale of the plot, independentiy of life stage and height on Fig. 5. Chthamalus steiiatus. Mean number of recruits (+SE, the shore (Table 6). spatialvanance in mortality was n = 12) at different heights above Mean Low Water Level (MLWL) at 2 localities in the northwest Mediterranean also large at the scale of the patch both for juvenile and adult barnacles living in low-shore habitats, and at the scale of the location for juvenile barnacles living low on the shore. There was no temporal vanabiiity in pat- shore ones (Q2= 10.7, df = 1, p < 0.01). Sirnilarly, SNK terns of mortality of barnacles, and spatial vanabiiity at tests within the Height X Time X Location interaction the scale of the site was negligible (Table 6). indicated that mortality was significantly greater in low-shore habitats in all compansons (Fig. 7). Similar results were obtained for adult barnacles. Mortality in DISCUSSION low-shore habitats ranked highest significantly more often than did mortality in high-shore habitats within In agreement with the preliminary observations the Height X Patch interaction (Anderson's test: Q2= 6, made at Livorno (Benedetti-Cecchi & Acunto unpubl. df = 1,p < 0.02).An example of this smali-scale spatial data), Chthamalus stellatus was significantly and con- variability in differences between heights is given for sistently larger higher up than lower down on the 1 location (Livorno), using data from Year 1 (Fig. 8). shore in the northwest Mediterranean. Similarly, the SNK tests within the Height X Time X Location interac- Cover of this species was greater in high-shore than Benedetti-Cecchi et al.: Population ecology of barnacles 165

Table 5. Chtharnalus stellatus. ANOVA of mean growth of juveniles and adults. Significant effects relevant to the interpretation of the results are indicated in bold. T: Time

Source of vanation df Juveniles Adults Denominator for F MS F P MS F P

T 1 0.0013 0.0 >0.9 0.1584 0.1 >0.8 TxL L 2 2.7341 - - 1.6454 - - No test Site (L) 6 0.1125 1.2 >0.4 0.0370 0.8 >0.5 T X Site (L) TxL 2 1.4242 14.7 <0.005 2.9332 65.8 <0.0002 T X Site (L) T X Site (L) 6 0.0971 1.2 >0.3 0.0446 6.9 <0.0002 Patch (T X Site (L)) Patch (T X Site (L)) 36 0.0829 5.1 <0.0001 0.0065 1.5 <0.05 Residual H 1 0.9582 - - 0.055lc 4.1 >0.05 H X T X Site (L) HxT 1 0.0281 0.2 >0.6 0.0127 2.4 >0.2 HxTxL HxL 2 0.1331 - 0.0O3OC 0.3 ~0.6 H X Site (L) HxTxL 2 0.1182a 3.6 <0.05 0.0054 0.4 >0.6 H X T X Site (L) H X Site (L) 6 0.0094 0.4 >0.8 0.0105 0.8 >0.6 H X T X Site (L) H X T X Site (L) 6 0.0243 0.7 >0.6 0.0134 2.1 >0.05 H X Patch (T X Site (L)) H X Patch (T X Site (L)) 36 0.0345 2.1 <0.0003 0.0063 1.4 >0.05 Residual Residual 432b 0.0170 0.0048 Cochran's C-test C = 0.072, p > 0.05 C = 0.080, p < 0.05 Transformation None Ln(x + 1)

aPooled term: H X T X Site (L) + H X Patch (T X Site (L));MS = 0.0331, 42 df b~ueto rnissing data, there were 413 df in the Residual for juveniles and 396 in the Residual for adults CAdenominator for these tests was provided after post-hoc elirnination of other terms in the model that were not significant at a = 0.25 (Winer et al. 1991)

low-shore habitats despite considerable spatial and Juveniles temporal vanation in patterns of abundance within each height. Successful recruitment was observed at different heights on the shore, when free space was provided. There were large differences in the density of recruits between the 2 locations considered. At Livorno, recruitment increased with height on the shore, but there was no clear pattern of significance in these effects. Few spat occurred in cleared plots at 0.20.0~~IElSl Punta Bianca, both in high-shore and low-shore habitats. Bare rock available for recruitment was greater Year 1 Year 2 high on the shore rather than lower down at all locations. There was no evidence for barnacles living higher up on the shore to grow at faster rates than those living lower down. The opposite pattern was observed for young barnacles in several cases. Mortal- ity rates were generally larger in low-shore than high- g 0.4 shore habitats for both young and adult barnacles. 0.2 These results emphasize the importance of pre-emp- 0.0 tion of space and mortality of juveniles in generating g Year l Year 2 patterns in the distribution and structure of populations of barnacles on rocky shores. C Livorno Despite a trend toward a larger number of recruits '.O1 higher on the shore at Livorno, differences among heights were not large enough to fully explain the dl 0.2:::M 1 Fig. 6. Chthamalus stellatus. Mean annual growth (+SE, n = 42 to 45 due to missing data) of juveniles in high-shore and 0.0 low-shore habitats at 3 locations in 2 consecutive years Year l Year 2 166 Mar Ecol Prog Ser 198: 157-170, 2000

Table 6.Estirnates of spatial and temporal variance in growth and mortality of barnacles

. . Response Life Height on 02plot 02Patch , 02site 02Locaiion 02Tie variable stage the shore

Growth Juvenile High Juvenile Low Adult High Adult Low Mortaiity Juvenile High Juvenile Low Adult High Adult Low

Table 7. Chthamalus steiiatus. ANOVA of mean mortality of juveniles and adults. Significant effects relevant to the interpretation of the results are indicated in bold

Source of variation df Juveniles Adults Denominator for F MS FP MS F P

T 1 5219.6 0.8 >0.4 2050.8 0.9 >0.4 TxL L 2 5851.2 - - 7316.6 - - No test Site (L) 6 1750.9 4.8 <0.05 1055.8 0.8 >0.5 T X Site (L) TxL 2 6731.4 18.6 <0.003 2186.3 1.7 >0.2 T X Site (L) T X Site (L) 6 361.2 0.5 >0.8 1308.7 0.9 >0.4 Patch (T X Site (L)) Patch (T X Site (L)) 36 799.5 1.6 <0.05 1406.8 4.5 <0.0001 Residual H 1 73909.9 - - 48460.2 - - No test HxT 1 2386.4 0.5 >0.5 51.7 0.1 >0.9 HxTxL HxL 2 2837.8 - - 1585.8 - No test HxTxL 2 4467.2 34.4 <0.0006 3485.4 6.4 <0.05 H X T X Site (L) H X Site (L) 6 254.4 2.0 >0.2 3874.2 7;l <0.05 H X T X Site (L) H X T X Site (L) 6 129.8 0.2 >0.9 545.4 0.7 >0.6 H X Patch (T X Site (L))

' H X Patch (T X Site (L)) 36 828.2 1.7 <0.01 841.1 2.7 <0.0001 Residual Residual 432a 552.0 355.2 Cochran's C-test C = 0.047,p > 0.05 C = 0.054,p > 0.05 Transformation None None aDue to rnissing data, there were 382 df in the Residual for juveniles and 379 in the Residual for adults

observed differences in percentage Cover. Thus, spa- Underwood 1979, Gaines & Roughgarden 1985, tial variation in the delivery of larvae across the verti- Chabot & Bourget 1988). Previous expenments in the cal gradient could explain only a small proportion of northwest Mediterranean have shown that the colo- variation in the distribution of young and adult barna- nization of barnacles is enhanced by the presence of cles in this rnicro-tidal System. This contrasts to obser- limpets, probably due to the indirect effects of these vations made in Systems where tides are more impor- grazers that prevented the monopolization of the sub- tant, where physical processes in the water column stratum by filamentous algae (Benedetti-Cecchi 2000). and local hydrodynamics near the substratum can gen- Resident organisms do not necessarily prevent or erate spatial vanation in density of larvae and then limit recruitment of barnacles. In high-shore habitats, influence the distnbution of adult animals (e.g. Gros- for example, the presence of ephemeral algae may berg 1982, Wethey 1986, Mullineaux & Butman 1991). increase the survlval of recently settled larvae, as In contrast, actual recruitment to natural populations in descnbed by Minchinton & Scheibling (1993) for the the northwest Mediterranean probably reflected the barnacle Semibalanus balanoides in Nova Scotia. variable effects of resident organisms at different Thus, organisms that have negative effects on barna- heights on the shore (barnacles high on the shore and cles in the more benign part of the habitat (low on the macroalgae lower down) on larvae of barnacles. Pre- shore), may exert positive effects in stressful conditions emption of the substratum is a common process limit- (high on the shore) (Bertness 1989, Stephens & Bert- ing settlement of larvae of these organisms (Denley & ness 1991). In the case of barnacles, positive effects Benedetti-Cecchi et ai.: Population ecology of barnacles 163

Juveniles Adults

A Site 1 High shore 0Low shore

60I01 B

20mlm-l0 Year 1 Year 2

B Site 2 801

20U0 Year 1 Year 2

C Livorno 801

20WO ~earl Year 2

Fig. 3. Chthamalus steiiatus. Mean annuai mortality (+SE, n = 33 to 45 due to missing data) of juveniles in high-shore and low-shore habitats at 3 locations in 2 consecutive years Fig. 8. Chthamalus stellatus. Mean annual mortality (+SE, n = 5) of adults in high-shore and low-shore habitats. Data are from 3 patches nested within 3 sites at Livorno in Year 1 rnay also result from chernical cues enhancing settlement of larvae close to adult individuals (Knight-Jones 1953, Crisp 1955, Wethey 1984, Chabot & Bourget known to be positively related to the concentration of 1988, Rairnondi 1988). These effects might explain the food in the water column and to the delivery of food by slightly larger number of recruits observed in plots water currents (Bertness et al. 1991, Leonard et al. high on the shore at Livorno, in comparison to quadrats 1998). Thus, supply of food should be larger low on the occurring lower down. These plots, in fact, were shore where barnacles are nearly continuously sub- cleared in a matrix of adult barnacles, while those in mersed than in high-shore habitats where, at least in low-shore habitats were surrounded by macroalgae. the Mediterranean, barnacles become submersed The results of this study indicate that high-shore mainly during storms. This is because the distribution habitats are not better than low-shore habitats in terms of barnacles high on the shore exceeds the mark of the of processes influencing the growth of Chthamalus high tide in the Mediterranean. In addition to the stellatus. Our analyses indicated that the annual delivery of food, growth of barnacles in high-shore increase in size of juvenile barnacles low on the shore habitats could be limited by physical Stress, although was similar, and in some cases larger, than that of bar- aggregation may reduce the effects of environmental nacles occurring in high-shore habitats. These results harshness in these organisms (Bertness 1989). Aggre- are similar to those found in other studies. For exam- gation, however, may result in elongated barnacles ple, Bertness et al. (1991) documented faster growth of with smali opercular or basal diameter (Barnes & Pow- Semibalanus balanoides in low-shore areas. A nurnber ell 1950, Bertness 1989), so that density and growth of processes could account for the differentes between may be negatively correlated across the vertical gradi- heights documented in this and our study. Growth is ent (when size is expressed as opercular or basal diam- 168 Mar Ecol Prog Ser

eter), providing an alternative explanation for the dif- these scales, however, hydrodynamic processes may ferences in patterns of growth between high-shore and also play a role (e.g. Eckrnan 1990). low-shore barnacles documented in the present and Variation in recruitment, growth and mortality may other studies. These processes are likely to be of minor have important ecological effects as these traits deter- importance in comparison to other factors. In fact, dif- mine the local density of populations and ultimately ferences in growth between high-shore and low-shore affect the patterns of intra- and inter-specific interac- barnacles are still evident when comparing the growth tions in assemblages. Several studies have explored of individuals maintained in isolation (Bertness et al. the consequences of variable recruitment in the sea 1991). (e.g. Underwood & Denley 1984, Roughgarden et al. In contrast to growth, processes affecting the mortal- 1987) and the implications of this variability for the ity of Chthamalus steiiatus in our study were more in- population ecology of barnacles have been exarnined tense low on the shore. These processes maintained in detail (Grosberg 1982, Hawkins & Hartnoll 1982, sparse, low-density populations relative to high-shore Caffey 1985, Gaines et al. 1985, Raimondi 1990, Pineda habitats. Studies on Semibalanus balanoides on rocky 1994, Ross & Underwood 1997, Hyder et al. 1998). In shores in Nova Scotia (Minchinton & Scheibling 1993), contrast, less has been done on spatial and temporal have shown greater post-recruitment mortality in low- variation in patterns of growth and mortality of these shore than high-shore intertidal habitats. Mortality oi organisms (but see Bertness et ai. i93ij. Understand- barnacles in low-shore environments is often a conse- ing variation in these traits is, however, important in quence of biological processes such as predation by order to make predictions about the effects of barna- whelks (Connell 1961b, Minchinton & Scheibling 1993) cles in assemblages. In particular, growth and final size and overgrowth by macroalgae, particularly encrusting strongly affect the capability of barnacles to compete algae (Denley & Underwood 1979, Bertness et al. 1983, for space with related taxa and with mobile animals 1991, Bertness 1989). Whelks were uncommon on our such as lirnpets (e.g. Connell 1961a,b,Farre11 1991). In shores, although they were locally abundant at some the northwest Mediterranean, dense populations of locations (e.g. at some sites in Capraia). Although we large barnacles can affect the establishment of foliose cannot discount predation as a processes influencing algae high on the shore (Benedetti-Cecchi et al. 1999); mortality of barnacles, it probably played a minor role they can also reduce the local density of lirnpets in in determining the observed patterns. In contrast, over- rnid-shore habitats, indirectly facilitating the establish- growth by algae was a likely mechanism. In several ment of ephemeral algae late in succession (Benedetti- cases we observed filamentous and encrusting algae Cecchi 2000). Therefore, changes in density of barna- over dead barnacles in low-shore habitats. Clearly, ex- cles can have both direct and indirect consequences in perirnents are needed to determine whether barnacles assemblages on our rocky coasts. are kiiled by algae or, alternatively, algae grow over In conclusion, the results of this study emphasize the barnacles once they are killed by other factors. importance of free space, mortality and recruitment as Our analysis on spatial variance at different scales factors affecting the structure of populations of barna- emphasized the importance of both regional and local cles in the northwest Mediterranean. Refinement of processes in regulating patterns of recruitment, this space-mortality-recruitment model will require growth and mortality of Chthamalus steiiatus in the additional research on the processes causing mortality northwest Mediterranean. Oceanographic conditions to low-shore individuals and on the causes of variabil- influencing the delivery of larvae and food rnight ity in these processes. More generally, our analysis account for greater recruitment and growth of barna- indicates that future investigations on the population cles at Livorno than at the other locations. The impor- ecology of Chthamalus stellatus should address the tance of these processes in generating large-scale vari- causes of variation at the regional (10s to 100s of km) ation in the demography and structure of populations and local (cm to m) spatial scales. of barnacles has been documented in other studies (Bertness et al. 1991, Sanford et al. 1994, Leonard et al. Acknowledgements. We sincerely thank Prof. A. Myers for 1998). In contrast, intra- and inter-specific interactions his help with the manuscript. The final version benefited from and the effects of heterogeneity of the substratum are comrnents by Tony Underwood and 3 anonymous referees. This study was supported by the EU under the MAS3-CT95- usually invoked as causal processes influencing pat- 0012 Programme (Eurorock)and by a grant from the Univer- terns at small spatial scales (Connell 1961a,b, Chabot sity of Pisa. & Bourget 1988, Lively et al. 1993). All these processes are potential candidates to explain the vanability in LITERATURE CITED recruitment, growth and mortaiity of barnacles observed in the present study, at scales of 10s of cm Barnes H, Poweil HT (1950) The development, general mor- (among plots) to a few metres (among patches). At phology, and subsequent eliminination of barnacle popu- Benedetti-Cecchi et al.: Popidation ecology of barnacles 169

lations after a heavy initial settlement. J Anirn Ecol 19: Hyder K, Johnson MP, Hawkins SJ, Gurney WSC (1998) Bar- 175-179 nacle demography: evidence for an existing model and Benedetti-Cecchi L (2000) Predicting direct and indirect in- spatial scales of variation. Mar Ecol Prog Ser 174:lOl-106 teractions dunng succession in a midlittoral rocky shore Knight-Jones EW (1953) Laboratory experiments on gregari- assemblage. Ecol Monogr 70:45-72 ousness during settling in Balanus balanoides and other Benedetti-Cecchi L, Airoldi L, Abbiati M, Cinelli F (1996) Esti- barnacles. J Exp Biol 30:584-598 mating the abundance of benthic invertebrates: a compar- Leonard GH, Levine JM, Schmidt PR, Bertness MD (1998) ison of procedures and variability between observers. Mar Flow-driven variation in intertidal community structure in Ecol Prog Ser 138:93-101 a Maine estuary. Ecology 79(4):1395-1411 Benedetti-Cecchi L, Menconi M, Cineili F (1999) Preemption Levin SA (1992) The problem of pattern and scale in ecology. of the substratum and the maintenance of spatial pattern Ecology 73:1943-1967 on a rocky shore in the northwest Mediterranean. Mar Lively CM, Raimondi PT, Delph LF (1993) Intertidal commu- Ecol Prog Ser 181:13-23 nity structure: space-time interactions in the Northern Bertness MD (1989) Positive and negative density-dependent Gdf of California. Ecology 74:162-173 mortahty and the population structure of Semibalanus bal- Menconi M, Benedetti-Cecchi L, Cinelli F (1999) Spatial and anoides in a sheltered bay habitat. Ecology 70:257-268 temporal variability in the distnbution of algae and inver- Bertness MD, Yund PO, Brown AF (1983) Snail grazing and tebrates on rocky shores in the northwest Mediterranean. the abundance of algal crusts on a sheltered New England J Exp Mar Biol Ecol233:l-23 rocky beach. J Exp Mar Biol Ecol71:147-164 Minchinton TE, Scheibling RE (1993) Free space availability Bertness MD, Gaines SD, Bermudez D, Sanford E (1991) and larval substratum selection as determinants of barna- Extreme spatial vanation in the growth and reproductive cle population structure in a developing rocky intertidal output of the acorn barnacle Semibalanus balanoides. Mar community. Mar Ecol Prog Ser 95:233-244 Ecol Prog Ser 75:91-100 Mullineaux LS, Butrnan CA (1991) Initial contact, exploration Caffey HM (1985) Spatial and temporal vanation in the settle- and attachment of barnacle (Balanus amphitrite) cypnds ment and recruitment of intertidal barnacles. Ecol Monogr settling in flow. Mar Biol 110:93-103 55:313-335 Pineda J (1994) Spatial and temporal patterns in barnacle set- Carlile DW, Skalski JR, Batker JE, Thomas JM, Cullinan V1 tlement rate along a southern California rocky shore. Mar (1989) Determination of ecological scale. Landscape Ecol Ecol Prog Ser 107:125-138 2:203-213 Raimondi PT (1988) Settlement cues and determination of the Chabot R, Bourget E (1988) Influence of substratum hetero- vertical Limit of an intertidal barnacle. Ecology 69:400-407 geneity and settled barnacle density on the settlement of Raimondi PT (1990) Patterns, mechanisms, consequences of cypris larvae. Mar Biol 97:45-56 variability in settlement and recruitrnent of an intertidal Conneil JH (1961a) The influence of interspecific competition barnacle. Ecol Monogr 60:283-309 and other factors on the distnbution of the barnacle Ross PM, Underwood AJ (1997) The distribution and abun- Chthamalus stellatus. Ecology 42:710-723 dance of barnacles in a mangrove forest. Aust J Ecol 22: Connell JH (1961b) The effect of competition, predation by 37-47 Thais lapillus and other factors on natural population of Rossi RE, Mulla DJ, Jurnel AG, Franz EH (1992)Geostatistical the barnacle Balanus balanoides. Ecol Monogr 31:61-104 tools for modeling and interpreting ecological spatial Cnsp DJ (1955)The behavior of barnacle cypnds in relation to dependence. Ecol Monogr 62:277-314 water movement over a surface. J Exp Biol 32:569-590 Roughgarden J, Iwasa Y, Baxter C (1985) Demographic the- Denley EJ, Undenvood AJ (1979) Experiments on factors ory for an Open marine population with space-lunited influencing settlement, survival, and growth of two spe- recruitment. Ecology 66:54-67 cies of barnacles in New South Wales. J Exp Mar Biol Ecol Roughgarden J, Gaines SD, Pacala S (1987) Supply side eco- 36:269-293 logy: the role of physical transport processes. In: Giiler P, Dethier MN, Graham ES, Cohen S, Tear LM (1993) Visual Gee J (eds) Organization of communities: past and pre- versus random-point percent Cover estimations: 'objective' sent. Proc Br Ecol Soc Symp, Aberystwyth, Wales (April is not always better. Mar Ecol Prog Ser 96:93-100 1986).Blackweil Scientific Publications, London, p 459-489 Eckman JE (1990) A model of passive settlement by plank- Roughgarden J, Gaines SD, Possingham H (1988) Recruit- tonic larvae onto bottoms of differing roughness. Mar Ecol ment dynamics in complex life cycles. Science 241: Prog Ser 353387-901 1460-1466 Farrell TM (1991) Models and mechanisms of succession: an Sanford E, Bermudez D, Bertness MD, Gaines SD (1994) Flow, exarnple from a rocky intertidal cornrnunity. Ecol Monogr food supply and acorn barnacle population dynamics. Mar 61:95-113 Ecol Prog Ser 104:49-62 Gaines SD, Roughgarden J (1985) Larval settlement rate: a Schneider DC (1994) Quantitative ecology: spatial and tem- leading determinant of structure in an ecological commu- poral scaling. Academic Press, San Diego nity of the marine intertidal Zone. Proc Natl Acad Sci USA Southward AJ (1987) Barnacle biology. AA Balkema, Rotter- 82:3707-37 11 dam Gaines SD, Brown S, Roughgarden J (1985) Spatial variation Stephens EG, Bertness MD (1991) Mussel facilitation of bar- in larval concentration as a cause of spatial variation in nacle suMval in a sheItered bay habitat. J Exp Mar Biol settlement for the barnacle, Balanus glandula. Oecologia Eco1145:33-48 67:267-272 Stephenson TA, Stephenson A (1949) The universal feature of Grosberg RK (1982) Intertidal zonation of barnacles: the influ- zonation between tide-marks on rocky coasts. J Ecol 37: ence of planktonic zonation of larvae on vertical distribu- 289-305 tion of adults. Ecology 63:894-899 Thrush SF, Pndmore RD, Hewitt JE (1994) Impacts on soft- Hawkins SJ, Hartnoil RG (1982) Settlement patterns of Semi- sediment macrofauna: the effects of spatial variation on balanus balanoides (L.) in the Isle of Man (1977-1981). temporal trends. Ecol Appl4:31-41 J Exp Mar Biol Ecol62:271-283 Underwood AJ (1997) Experiments in ecology: their logical 170 Mar Ecol Prog Ser 198: 157-170,2000

design and interpretation using analysis of variance. Cam- ties: conceptual issues and the evidente. Princeton Uni- bndge University Press, Cambridge. versity Press, Princeton, NJ, p 151-180 Underwood AJ, Chapman MJ (1996) Scales of spatial patterns Wethey DS (1984) Spatial patterns in barnacle settlement: day of distribution of intertidal invertebrates. Oecologia 107: to day changes during the settlement season. J Mar Biol 212-224 Assoc UK 64:687-698 Underwood AJ, Denley EJ (1984) Paradigms, explanations, Wethey DS (1986) Ranking of settlement cues by barnacle lar- and generalizations in models for the structure of inter- vae: infiuence of surface contour. Bull Mar Sci 39:393-400 tidal communities on rocky shores. In: Strong DR, Sim- Wier BJ, Brown DR, Michelis KM (1991)-Statisticalprinciples berloff D, Abele LG, Thistle AB (eds) Ecological communi- in experimental design, 3rd edn. McGraw-Hiü, New York

Editorial responsibility: Otto Ki~e(Editor), - Submitted: September 6, 1999; Accepted: December 1, 1999 Oldendorf/Luhe, Germany Proofs received from author(s): May 8, 2000